Open Access Highly Accessed Research article

EMT is the dominant program in human colon cancer

Andre Loboda1, Michael V Nebozhyn1, James W Watters1, Carolyne A Buser3, Peter Martin Shaw2, Pearl S Huang3, Laura Van't Veer7, Rob AEM Tollenaar8, David B Jackson6, Deepak Agrawal5, Hongyue Dai4 and Timothy J Yeatman5*

Author Affiliations

1 Merck, Sharp and Dohme, P.O. Box 4, 770 Sumneytown Pike, Building 53, West Point, PA 19486, USA

2 Merck, Sharp and Dohme, P.O. Box 4, 770 Sumneytown Pike, WP53B-120, West Point, PA 19486, USA

3 Merck Research Laboratories, UG-4C74, 351 N. Sumneytown Pike, North Wales, PA 19454-2505, USA

4 Merck, Sharp and Dohme, 33 Avenue Louis Pasteur, Boston, MA 02115, USA

5 Moffitt Cancer Center, 12902 Magnolia Drive, Tampa, FL 33612, USA

6 LIFE Biosystems GmbH, Poststrasse 34, D-69115 Heidelberg, Germany

7 Diagnostic Oncology, The Netherlands Cancer Institute, Plesmanlaan 121, 1066 CX, Amsterdam, The Netherlands

8 Leiden University Medical Center, PO Box 9600 2300 RC Leiden, The Netherlands

For all author emails, please log on.

BMC Medical Genomics 2011, 4:9  doi:10.1186/1755-8794-4-9

Published: 20 January 2011

Additional files

Additional file 1:

Ingenuity/GO Analysis produced multiple functional categories for PC1 without bringing clarity to the underlying biology. The table lists top functional gene groups in terms of significance for enrichment of genes from PC1 signature. Both, significance of enrichment p-value,(based on hyper geometric distribution) and Bonferroni-type correction e-value (to account for multiple testing). Gene sets from Ingenuity, KEGG, and GeneGO were included in the analysis.

Format: PDF Size: 35KB Download file

This file can be viewed with: Adobe Acrobat Reader

Open Data

Additional file 2:

EMT signature was derived by comparing gene expression of cell lines sorted into epithelial or mesenchymal like groups based on CDH1 and VIM expression (see Additional Figure 1). The top 200 up and down probes found most significant by ANOVA (P < 0.001) were selected to represent the EMT signature. The EMT signature contains known EMT drivers such as ZEB1 and ZEB2, TCF4, AXL. It also contains markers such as CDH1, CDH3 for epithelial phenotype and VIM, CDH2 and CDH4 for the mesenchymal phenotype.

Format: PDF Size: 24KB Download file

This file can be viewed with: Adobe Acrobat Reader

Open Data

Additional file 3:

miRNA correlation to EMT and RAS signature scores on mean-centered data. Pearson correlation coefficient and the associated p-value are provided.

Format: PDF Size: 17KB Download file

This file can be viewed with: Adobe Acrobat Reader

Open Data

Additional file 4:

Centered abundances for 49 tumors × 416 MiR detectors.

Format: PDF Size: 41KB Download file

This file can be viewed with: Adobe Acrobat Reader

Open Data

Additional file 5:

Out of ~300 signatures tested, EMT was the most significantly associated with PC1 in colon (P < 10-135). More importantly, the up and down arms of the EMT signature were directionally correlated with PC1 (P < 10-16, Fisher Exact Test). See Additional File 2 for list of genes.

Format: PDF Size: 44KB Download file

This file can be viewed with: Adobe Acrobat Reader

Open Data

Additional file 6:

Derivation of the EMT signature used to clarify the biology characterizing PC1. The EMT signature was derived from a global gene expression analysis of 93 lung cancer cell lines first segregated by differential CDH1 and VIM expression. Right panel shows the relationship between EMT signature score and CDH1 probe intensities, the left panel shows the EMT signature score vs. VIM probe intensity. EMT signature is observed to be positively correlated to VIM and anticorrelated to CDH1.

Format: PDF Size: 40KB Download file

This file can be viewed with: Adobe Acrobat Reader

Open Data

Additional file 7:

Waterfall plot of recurrence prediction of PC1 for the MCC colon dataset shows more recurrences with high signature scores than with low signature scores; similarly there fewer recurrences with low signature scores than with high signature scores.

Format: PDF Size: 16KB Download file

This file can be viewed with: Adobe Acrobat Reader

Open Data

Additional file 8:

Hierarchical cluster analysis showing expression of key genes (red and blue) and gene signatures (black) in the EMT signature for colorectal tumors. Genes positively correlated with the EMT signature are shown in red and genes negatively correlated with the EMT signature are shown in blue.

Format: PDF Size: 20KB Download file

This file can be viewed with: Adobe Acrobat Reader

Open Data

Additional file 9:

Hierarchical cluster analysis showing expression of key genes (red and blue) and gene signatures (black) in the EMT signature for lung tumors. Genes positively correlated with the EMT signature are shown in red and genes negatively correlated with the EMT signature are shown in blue.

Format: PDF Size: 22KB Download file

This file can be viewed with: Adobe Acrobat Reader

Open Data

Additional file 10:

Hierarchical cluster analysis showing expression of key genes (red and blue) and gene signatures (black) in the EMT signature for pancreatic tumors. Genes positively correlated with the EMT signature are shown in red and genes negatively correlated with the EMT signature are shown in blue.

Format: PDF Size: 18KB Download file

This file can be viewed with: Adobe Acrobat Reader

Open Data

Additional file 11:

Waterfall and boxplot analysis's shows a differential EMT score for colon < lung < pancreas following normalization across all samples.

Format: PDF Size: 31KB Download file

This file can be viewed with: Adobe Acrobat Reader

Open Data

Additional file 12:

Top 5000 most variable genes (columns) on Colon cell lines (rows) sorted by PC1. PC1 is observed to be positively correlated to EMT signature score and anti-correlated to RAS signature score. Genes are clustered using Pearson correlation distance metric and Ward linkage. Heatmap shows mean-centered probe intensities.

Format: PDF Size: 84KB Download file

This file can be viewed with: Adobe Acrobat Reader

Open Data

Additional file 13:

PC1 predicts recurrence in stages 2 and 3 of colon cancer. Data is shown for MCC dataset.

Format: PDF Size: 62KB Download file

This file can be viewed with: Adobe Acrobat Reader

Open Data

Additional file 14:

Covariance matrices showing the relationship of PC1 to the same endpoints as shown in Figure 4ausing (a) independent colon dataset [21](b) EXPO dataset, (c) NKI dataset.

Format: PDF Size: 41KB Download file

This file can be viewed with: Adobe Acrobat Reader

Open Data

Additional file 15:

EMT signature proposed in this paper is predictive of recurrence in stage 2 and stage 3 MCC tumors.

Format: PDF Size: 16KB Download file

This file can be viewed with: Adobe Acrobat Reader

Open Data

Additional file 16:

EMT signature proposed in this paper is predictive of recurrence when applied to all tumor samples in MCC data set.

Format: PDF Size: 32KB Download file

This file can be viewed with: Adobe Acrobat Reader

Open Data